mlsanityData to 14 Oct 2026

TaQ-DiT: Time-aware Quantization for Diffusion Transformers

arXiv 2411.14172 ↗First posted 21 Nov 2024eess.IVcited by 2

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Transformer-based diffusion models, dubbed Diffusion Transformers (DiTs), have achieved state-of-the-art performance in image and video generation tasks. However, their large model size and slow inference speed limit their practical applications, calling for model compression methods such as quantization. Unfortunately, existing DiT quantization methods overlook (1) the impact of reconstruction and (2) the varying quantization sensitivities across different layers, which hinder their achievable performance. To tackle these issues, we propose innovative time-aware quantization for DiTs (TaQ-DiT). Specifically, (1) we observe a non-convergence issue when reconstructing weights and activations separately during quantization and introduce a joint reconstruction method to resolve this problem. (2) We discover that Post-GELU activations are particularly sensitive to quantization due to their significant variability across different denoising steps as well as extreme asymmetries and variations within each step. To address this, we propose time-variance-aware transformations to facilitate more effective quantization. Experimental results show that when quantizing DiTs' weights to 4-bit and activations to 8-bit (W4A8), our method significantly surpasses previous quantization methods.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 24 of its 30 references that are in the map. First posted on arXiv 21 Nov 2024.

Builds on

24 of 30 references.

Closest ancestors · 12

Same direction, newest first

  1. 2024DiTAS: Quantizing Diffusion Transformers via Enhanced Activation Smoothing12 Sep
  2. 2024Q-DiT: Accurate Post-Training Quantization for Diffusion Transformers25 Jun
  3. 2024PTQ4DiT: Post-training Quantization for Diffusion Transformers25 May
  4. 2023EfficientDM: Efficient Quantization-Aware Fine-Tuning of Low-Bit Diffusion Models5 Oct
  5. 2023Towards Accurate Post-training Quantization for Diffusion Models30 May
  6. 2023PTQD: Accurate Post-Training Quantization for Diffusion Models18 May
  7. 2023Q-Diffusion: Quantizing Diffusion Models8 Feb
  8. 2022Scalable Diffusion Models with Transformers19 Dec
  9. 2022Post-training Quantization on Diffusion Models28 Nov
  10. 2022Diffusion Models in Vision: A Survey10 Sep
  11. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
  12. 2020Denoising Diffusion Probabilistic Models19 Jun

Other references · 12

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  2. 2022SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Modelscited by 1,386
  3. 2022ZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale Transformerscited by 526
  4. 2021BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstructioncited by 479
  5. 2024Sora: A Review on Background, Technology, Limitations, and Opportunities of Large Vision Modelscited by 477
  6. 2022QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantizationcited by 180
  7. 2021PTQ4ViT: Post-training quantization for vision transformers with twin uniform quantizationcited by 179
  8. 2021FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformercited by 151
  9. 2022RepQ-ViT: Scale Reparameterization for Post-Training Quantization of Vision Transformerscited by 137
  10. 2022Mixed-Precision Neural Network Quantization via Learned Layer-wise Importancecited by 54
  11. 2023Towards High-quality HDR Deghosting with Conditional Diffusion Modelscited by 33
  12. 2021SDTP: Semantic-aware Decoupled Transformer Pyramid for Dense Image Predictioncited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 1; 2025: 0; 2026: 1.

2 papers.

  1. 2024Efficiency Meets Fidelity: A Novel Quantization Framework for Stable Diffusioncited by 3
  2. 2026Post-training Quantization for Hybrid Iterative Generative Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Post-Training Quantization for Diffusion Transformer via Hierarchical Timestep Grouping
  2. 2026AdaTSQ: Pushing the Pareto Frontier of Diffusion Transformers via Temporal-Sensitivity Quantization
  3. 2024An Analysis on Quantizing Diffusion Transformers
  4. 2025TQ-DiT: Efficient Time-Aware Quantization for Diffusion Transformers
  5. 2025LRQ-DiT: Log-Rotation Post-Training Quantization of Diffusion Transformers for Image and Video Generation
  6. 2025TreeQ: Pushing the Quantization Boundary of Diffusion Transformer via Tree-Structured Mixed-Precision Search
  7. 2024ViDiT-Q: Efficient and Accurate Quantization of Diffusion Transformers for Image and Video Generation
  8. 20266Bit-Diffusion: Inference-Time Mixed-Precision Quantization for Video Diffusion Models